godot-docs/contributing/documentation/docs_writing_guidelines.rst

.. _doc_docs_writing_guidelines:

Writing guidelines
==================

The Godot community is rich and international. Users come from all
around the world. Some of them are young, and many aren't native English
speakers. That's why we must all write using a clear and a common
language. For the class reference, the goal is to make it easy to read
for everyone and precise.

In summary, always try to:

1. Use the active voice
2. Use precise action verbs
3. Avoid verbs that end in -ing
4. Remove unnecessary adverbs and adjectives.
5. Ban these 8 words: obvious, simple, basic, easy, actual, just, clear, and however
6. Use explicit references
7. Use 's to show possession
8. Use the Oxford comma

There are 3 rules to describe classes:

1. Give an overview of the node in the brief description
2. Mention what methods return if it's useful
3. Use "if true" to describe booleans

.. note::

    A technical writer's job is to pack as much information as possible into
    the smallest and clearest sentences possible. These guidelines will help
    you work towards that goal.

.. seealso::

    See the :ref:`content guidelines <doc_content_guidelines>` for information
    on the types of documentation you can write in the official documentation.

7 rules for clear English
-------------------------

Use the active voice
~~~~~~~~~~~~~~~~~~~~

Use the active voice when possible. Take the classes, methods, and
constants you describe as the subject. It's natural to write using the
passive voice, but it's harder to read and produces longer sentences.

.. highlight:: none

Passive:

::

    The man **was bitten** by the dog.

Active:

::

    The dog bit the man.

**Don't** use the passive voice:

::

    void edit_set_pivot ( Vector2 pivot )
    [...] This method **is implemented** only in some nodes that inherit Node2D.

**Do** use the node's name as a noun:

::

    void edit_set_pivot ( Vector2 pivot )
    [...] Only some Node2Ds **implement** this method.

Use precise action verbs
~~~~~~~~~~~~~~~~~~~~~~~~

Favor precise yet common verbs over generic ones like ``make``, ``set``,
and any expression you can replace with a single word.

**Don't** repeat the method's name. It already states it sets the pivot
value to a new one:

::

    void edit_set_pivot ( Vector2 pivot )
    Set the pivot position of the 2D node to [code]pivot[/code] value. [...]

**Do** explain what's the consequence of this "set": use precise verbs
like ``place``, ``position``, ``rotate``, ``fade``, etc.

::

    void edit_set_pivot ( Vector2 pivot )
    Position the node's pivot to the [code]pivot[/code] value. [...]

Avoid verbs that end in -ing
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

The progressive forms describe continuous actions. E.g. "is calling",
"is moving".

**Don't** use the progressive form for instant changes.

::

    Vector2 move ( Vector2 rel_vec )
    Move the body in the given direction, **stopping** if there is an obstacle. [...]

**Do** use simple present, past, or future.

::

    Vector2 move ( Vector2 rel_vec )
    Moves the body in the vector's direction. The body **stops** if it collides with an obstacle. [...]

Exception: If the subject is not clear, replacing "ing" verbs is not an
improvement. For example, in the previous sentence, "it replaces"
would not make much sense where "replacing" currently is.

You may use the progressive tense to describe actions that are
continuous in time. Anything like animation or coroutines.

.. tip::

    Verbs can turn into adjectival nouns with -ing. This is not a
    conjugation, so you may use them: ``the remaining movement``,
    ``the missing file``, etc.

Remove unnecessary adverbs and adjectives
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Write as few adjectives and adverbs as possible. Only use them if they
add key information to the description.

**Don't** use redundant or meaningless adverbs. Words that lengthen the
documentation but don't add any information:

::

    **Basically** a big texture [...]

**Do** write short sentences in a simple, descriptive language:

::

    A big texture [...]

Ban these 8 words
~~~~~~~~~~~~~~~~~

**Don't** ever use these 8 banned words:

1. obvious
2. simple
3. basic
4. easy
5. actual
6. just
7. clear
8. however (some uses)

Game creation and programming aren't simple, and nothing's easy to
someone learning to use the API for the first time. Other words in the
list, like ``just`` or ``actual`` won't add any info to the sentence.
Don't use corresponding adverbs either: obviously, simply, basically,
easily, actually, clearly.

**Don't** example. The banned words lengthen the description and take
attention away from the most important info:

::

    **TextureRect**
    Control frame that **simply** draws an assigned texture. It can stretch or not. It's a **simple** way to **just** show an image in a UI.

**Do** remove them:

::

    **TextureRect**
    [Control] node that displays a texture. The texture can stretch to the node's bounding box or stay in the center. Useful to display sprites in your UIs.

"Simple" never helps. Remember, for other users, anything could be
complex or frustrate them. There's nothing like a good old *it's simple*
to make you cringe. Here's the old brief description, the first sentence
on the Timer node's page:

::

    **Timer**
    A **simple** Timer node.

**Do** explain what the node does instead:

::

    **Timer**
    Calls a function of your choice after a certain duration.

**Don't** use "basic", it is too vague:

::

    **Vector3**
    Vector class, which performs **basic** 3D vector math operations.

**Do** use the brief description to offer an overview of the node:

::

    **Vector3**
    Provides essential math functions to manipulate 3D vectors: cross product, normalize, rotate, etc.

Use explicit references
~~~~~~~~~~~~~~~~~~~~~~~

Favor explicit references over implicit ones.

**Don't** use words like "the former", "the latter", etc. They're not
the most common in English, and they require you to check the reference.

::

    [code]w[/code] and [code]h[/code] define right and bottom margins. The **latter** two resize the texture so it fits in the defined margin.

**Do** repeat words. They remove all ambiguity:

::

    [code]w[/code] and [code]h[/code] define right and bottom margins. **[code]w[/code] and [code]h[/code]** resize the texture so it fits the margin.

If you need to repeat the same variable name 3 or 4 times, you probably
need to rephrase your description.

Use 's to show possession
~~~~~~~~~~~~~~~~~~~~~~~~~

Avoid "The milk **of** the cow". It feels unnatural in English. Write "The cow's
milk" instead.

**Don't** write "of the X":

::

    The region **of the AtlasTexture that is** used.

**Do** use ``'s``. It lets you put the main subject at the start of the
sentence, and keep it short:

::

    The **AtlasTexture's** used region.

Use the Oxford comma to enumerate anything
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

From the Oxford dictionary:

    The 'Oxford comma' is an optional comma before the word 'and' at the end of a list:
    *We sell books, videos, and magazines.*

    [...] Not all writers and publishers use it, but it can clarify the meaning of a sentence when the items in a list are not single words:
    *These items are available in black and white, red and yellow, and blue and green.*

**Don't** leave the last element of a list without a comma:

::

    Create a CharacterBody2D node, a CollisionShape2D node and a sprite node.

**Do** add a comma before `and` or `or`, for the last
element of a list with more than two elements.

::

    Create a CharacterBody2D node, a CollisionShape2D node, and a sprite node.


How to write methods and classes
--------------------------------

Dynamic vs static typing
~~~~~~~~~~~~~~~~~~~~~~~~

The code examples in the documentation should follow a consistent style not to
confuse users. As static type hints are an optional feature of GDScript, we
chose to stick to writing dynamic code. This leads to writing GDScript that is
concise and accessible.

The exception is topics that explain static typing concepts to users.

**Don't** add a type hint with a colon or by casting:

::

    const MainAttack := preload("res://fire_attack.gd")
    var hit_points := 5
    var name: String = "Bob"
    var body_sprite := $Sprite2D as Sprite2D


**Do** write constants and variables with dynamic typing:

::

    const MainAttack = preload("res://fire_attack.gd")
    var hit_points = 5
    var name = "Bob"
    var body_sprite = $Sprite2D


**Don't** write functions with inferred arguments or return types:

::

    func choose(arguments: PackedStringArray) -> String:
        # Chooses one of the arguments from array with equal chances
        randomize()
        var size := arguments.size()
        var choice: int = randi() % size
        return arguments[choice]

**Do** write functions using dynamic typing:

::

    func choose(arguments):
        # Chooses one of the arguments from array with equal chances
        randomize()
        var size = arguments.size()
        var choice = randi() % size
        return arguments[choice]

Use real-world code examples where appropriate
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Real-world examples are more accessible to beginners than abstract ``foos`` and
``bars``. You can also copy them directly from your game projects, ensuring that
any code snippet compiles without errors.

Writing ``var speed = 10`` rather than ``var my_var = 10`` allows beginners to
understand code better. It gives them a frame of reference as to where they
could use the code snippets in a live project.

**Don't** write made-up examples:

.. code-block:: gdscript

    @onready var a = preload("res://MyPath")
    @onready var my_node = $MyNode


    func foo():
        # Do stuff

**Do** write concrete examples:

.. code-block:: gdscript

    @onready var sfx_player_gun = preload("res://Assets/Sound/SFXPlayerGun.ogg")
    @onready var audio_player = $Audio/AudioStreamPlayer


    func play_shooting_sound():
        audio_player.stream = sfx_player_gun
        audio_player.play()

Of course, there are times when using real-world examples is impractical. In
those situations, you should still avoid using names such as ``my_var``,
``foo()`` or ``my_func()`` and consider more meaningful names for your examples.

Give an overview of the node in the brief description
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

The brief description is the reference's most important sentence. It's
the user's first contact with a node:

1. It's the only description in the "Create New Node" dialog.
2. It's at the top of every page in the reference

The brief description should explain the node's role and its
functionality, in up to 200 characters.

**Don't** write tiny and vague summaries:

::

    **Node2D**
    Base node for 2D system.

**Do** give an overview of the node's functionality:

::

    **Node2D**
    A 2D game object, inherited by all 2D-related nodes. Has a position, rotation, scale, and Z index.

Use the node's full description to provide more information, and a code
example, if possible.

Mention what methods return if it's useful
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Some methods return important values. Describe them at the end of the
description, ideally on a new line. No need to mention the return values
for any method whose name starts with ``set`` or ``get``.

**Don't** use the passive voice:

::

    Vector2 move ( Vector2 rel_vec )
    [...] The returned vector is how much movement was remaining before being stopped.

**Do** always use "Returns".

::

    Vector2 move ( Vector2 rel_vec )
    [...] Returns the remaining movement before the body was stopped.

Notice the exception to the "direct voice" rule: with the move method,
an external collider can influence the method and the body that calls
``move``. In this case, you can use the passive voice.

Use "if true" to describe booleans
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

For boolean member variables, always use ``if true`` and/or
``if false``, to stay explicit. ``Controls whether or not`` may be
ambiguous and won't work for every member variable.

Also, surround boolean values, variable names and methods with ``[code][/code]``.

**Do** start with "if true":

::

    Timer.autostart
    If [code]true[/code], the timer will automatically start when entering the scene tree.


Use ``[code]`` around arguments
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

In the class reference, always surround arguments with ``[code][/code]``. In the
documentation and in Godot, it will display like ``this``. When you edit XML
files in the Godot repository, replace existing arguments written like 'this' or
\`this\` with ``[code]this[/code]``.


Common vocabulary to use in Godot's documentation
-------------------------------------------------

The developers chose some specific words to refer to areas of the
interface. They're used in the sources, in the documentation, and you
should always use them instead of synonyms, so the users know what
you're talking about.

.. figure:: img/editor-vocabulary-overview.png
   :alt: Overview of the interface and common vocabulary

   Overview of the interface and common vocabulary

In the top left corner of the editor lie the ``main menus``. In the
center, the buttons change the ``workspace``. And together the buttons
in the top right are the ``playtest buttons``. The area in the center,
that displays the 2D or the 3D space, is the ``viewport``. At its top,
you find a list of ``tools`` inside the ``toolbar``.

The tabs or dockable panels on either side of the viewport are
``docks``. You have the ``FileSystem dock``, the ``Scene dock`` that
contains your scene tree, the ``Import dock``, the ``Node dock``, and
the ``Inspector`` or ``Inspector dock``. With the default layout you may
call the tabbed docks ``tabs``: the ``Scene tab``, the ``Node tab``...

The Animation, Debugger, etc. at the bottom of the viewport are
``panels``. Together they make up the ``bottom panels``.

Foldable areas of the Inspector are ``sections``. The node's parent
class names, which you can't fold, are ``Classes`` e.g. the
``CharacterBody2D class``. And individual lines with key-value pairs are
``properties``. E.g. ``position`` or ``modulate color`` are both
``properties``.

Keyboard shortcut guidelines
----------------------------

Keyboard and mouse shortcuts should make use of the ``:kbd:`` tag, which allows
shortcuts to stand out from the rest of the text and inline code. Use the
compact form for modifier keys (:kbd:`Ctrl`/:kbd:`Cmd`) instead of their spelled
out form (:kbd:`Control`/:kbd:`Command`). For combinations, use the ``+`` symbol
with a space on either side of the symbol.

Make sure to mention shortcuts that differ on macOS compared to other platforms.
You can find a list of all shortcuts, including what they are on macOS, on
:ref:`this page <doc_default_key_mapping>`.

Try to integrate the shortcut into sentences the best you can. Here are some
examples with the ``:kbd:`` tag left as-is for better visibility:

- Press ``:kbd:`Ctrl + Alt + T``` to toggle the panel (``:kbd:`Opt + Cmd + T``` on macOS).
- Press ``:kbd:`Space``` and hold the left mouse button to pan in the 2D editor.
- Press ``:kbd:`Shift + Up Arrow``` to move the node upwards by 8 pixels.


Manual style guidelines
-----------------------

Follow these formatting and style guidelines when writing the manual.

Use your best judgement. If you can write more clearly by breaking one of these
guidelines, please do! But remember that the guidelines exist for a reason.

.. note:: In many cases, the manual does not follow these guidelines. If you are
    already making changes to a paragraph or section of the docs, update it to
    follow these standards. Avoid making unrelated changes that *only* update style,
    since every change will require the paragraph to be re-translated.

Text styles
~~~~~~~~~~~

There are a few styles that the manual uses.

+---------------------+--------------------------+------------------------------------------------------------------------+
| Style               | RST formatting           | Typical usage                                                          |
+=====================+==========================+========================================================================+
| Plaintext           | ``text``                 | Used for most text.                                                    |
+---------------------+--------------------------+------------------------------------------------------------------------+
| *Italics*           | ``*text*``               | Used for emphasis. Used for introducing new terms.                     |
+---------------------+--------------------------+------------------------------------------------------------------------+
| **Bold**            | ``**text**``             | Used for emphasis, and for editor UI like menus and windows.           |
|                     |                          |                                                                        |
+---------------------+--------------------------+------------------------------------------------------------------------+
| ``Code``            | `` text ``               | Used for variable names, literal values, and code snippets. ``code`` is|
|                     |                          | used in many cases where you would use "quoted plaintext" in typical   |
|                     |                          | English.                                                               |
+---------------------+--------------------------+------------------------------------------------------------------------+
| "Quotes"            | ``"text"``               | Used for some literal or quoted values. In many cases, another         |
|                     |                          | style is preferred.                                                    |
+---------------------+--------------------------+------------------------------------------------------------------------+

Emphasis
~~~~~~~~

Use either **bold style** or *italic style* to emphasize words or sentences.
In most cases, either **bold** or *italics* is fine. Use whichever seems best,
or whatever the page already uses. 

Prefer using **bold style** for simple emphasis.

    - Do **not** close the window without saving first.

Use *italic style* or to emphasize one word in the context of a sentence.

    - You can *add* a node to the scene (but you can't connect one).
    - You can add a *node* to the scene (but you can't add a resource).
    - You can add a node to the *scene* (but you can't add one to a resource).

Use *italic style* when introducing new technical terms. **Bold style**
is fine too.

    - Godot uses *nodes* with *scripts* in a *scene tree*.
    - Godot uses **nodes** with **scripts** in a **scene tree**.

Literals
~~~~~~~~

Use ``code style`` for literal values. Literals include:

    - Integer or ``int`` literals like ``0``, ``-2``, or ``100``
    - Float literals like ``0.0``, ``0.5``, ``-2.0``, or ``100.0``
    - Vector literals like ``(0.0, 0.0)``, ``(0.5, -0.5, 0.5)``, or ``(1.0, 2.0, 3.0, 4.0)``.

Classes, properties, and methods
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Link to classes the first time that you mention them in a page. After the first
mention, use ``code style``. For common classes, like ``Node``, ``Control``, or
``Viewport``, you can also use plaintext.

Link to class members (properties, methods, enums, and constants) the first time
that you mention them in a page. After the first mention, use ``code style``. If
the class member is very common, like a Node2D's ``position``, you don't have to
link.

When discussing properties in the context of the inspector, use **bold style**
instead.

Editor UI
~~~~~~~~~

Use **bold style** for editor UI, including window titles, menus, buttons, input
fields, inspector properties, and inspector sections. Use the exact
capitalization that the editor uses.

    - Open the **Editor Settings** window.
    - Press the **Confirm** button.
    - Change the node's **Transform > Position** property to ``(0, 0)``.
    - In the **Project Settings** window, enable the **Advanced Settings** toggle.

Use **Bold > With > Separators** when describing sequence of menus that the
reader must navigate. Use ``>`` as a separator. You can omit ellipses in menu names.

    - In **Project > Project Settings > Input Map**, add a new input action.
    - Select **Scene > Export As... > MeshLibrary...**.
    - Select **Scene > Export As > MeshLibrary**.

.. note:: Sometimes, ``->`` or ``→`` is used as a separator. This is nonstandard.
    Replace it with ``>`` if you are already making changes to a section.

Project settings
~~~~~~~~~~~~~~~~

Link to individual project settings. Either include the section and subsection
in the link itself, or include the section and subsection separately from the
link. Since long links are not split into multiple lines when the page is
rendered, prefer splitting the setting name and the section when the link is long.

    - Set the :ref:`Application > Run > Max FPS<class_ProjectSettings_property_application/run/max_fps>` setting to ``60``.
    - In the project settings under **Application > Run**, set :ref:`Max FPS<class_ProjectSettings_property_application/run/max_fps>` to ``60``.
    - In **Project Settings > Application > Run**, set :ref:`Max FPS<class_ProjectSettings_property_application/run/max_fps>` to ``60``.

Manually wrapping lines
~~~~~~~~~~~~~~~~~~~~~~~

In the manual, lines must be manually wrapped to no more than 80-100 characters
per line. However, links must not be split into multiple lines, and can exceed
100 characters. Tables can also exceed 100 characters.

When making small changes, you don't need to manually re-wrap the whole paragraph,
as long as the lines don't exceed 100 characters.

**Bad:** Line length exceeds 100 characters:

.. code-block::

    The best thing to do is to wrap lines to under 80 characters per line. Wrapping to around 80-90 characters per line is also fine.
    If your lines exceed 100 characters, you definitely need to add a newline! Don't forget to remove trailing whitespace when you do.

**Good:** Lines are wrapped to 80-90 characters:

.. code-block::

    The best thing to do is to wrap lines to under 80 characters per line. Wrapping to
    around 80-90 characters per line is also fine. If your lines exceed 100 characters, you
    definitely need to add a newline! Don't forget to remove trailing whitespace when you do.

**Best:** Lines are wrapped to under 80 characters:

.. code-block::

    The best thing to do is to wrap lines to under 80 characters per line. Wrapping
    to around 80-90 characters per line is also fine. If your lines exceed 100
    characters, you definitely need to add a newline! Don't forget to remove
    trailing whitespace when you do.

.. tip:: In most text editors, you can add a vertical guide or "ruler" at 80
    characters. For example, in Visual Studio Code, you can add the following to
    your ``settings.json`` to add rulers at 80 and 100 characters:

    .. code:: json

        "editor.rulers": [80,100], 

When to refer to a specific Godot version
-----------------------------------------

Most of the time, the class reference and the manual should not specify the first
version in which a feature is added. This is because the documentation describes
the *current* features of the engine. Documentation will be read and maintained
for many versions after it is initially written, and a reference to a first supported
version is only relevant for a few versions after a feature is added. After that,
it becomes historical trivia best left to a dedicated changelog.

Follow these guidelines for when to refer to a specific Godot version:

- If a feature was added in the current major version (4.x), **you can specify**
  the feature is new in 4.x.
- If a feature or default approach to a problem was changed between major versions
  (3.x -> 4.x), describe the current feature in the main body of the page, and
  optionally add a brief sentence or note block to compare 3.x and 4.x.
- If a large feature is added in a 4.x minor version, **you can specify** the minor
  version when it was added. Large features have a whole page or large section of
  documentation. In many cases it should still be avoided, since it's only relevant
  for the next few minor versions.
- If a small feature is added in a 4.x minor version, **do not specify** the minor
  version when it was added. Small features have only a short section of
  documentation, or are minor additions to existing features.
- If the default approach to a problem is changed in a 4.x minor version, **do
  specify** the minor version in which a new default approach was added. For example,
  the change from ``TileMap`` to ``TileMapLayer`` in 4.3.
- If a feature was added in a 3.x major or minor version, **do not specify** when 
  the feature was added. These features are old enough that the exact version
  in which they were added is not relevant.